Improving greenhouse gas emissions intensities of subtropical and tropical beef farming systems using Leucaena leucocephala

Agricultural Systems 136 (2015) 138–146

Contents lists available at ScienceDirect

Agricultural Systems j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / a g s y

Improving greenhouse gas emissions intensities of subtropical and tropical beef farming systems using Leucaena leucocephala Matthew T. Harrison a,*, Chris McSweeney b, Nigel W. Tomkins c, Richard J. Eckard d a

Tasmanian Institute of Agriculture, University of Tasmania, Tas. 7320, Australia CSIRO Agriculture, Queensland BioScience Precinct, St Lucia, Brisbane, Qld 4067, Australia CSIRO Agriculture, ATSIP, James Cook University, Townsville, Qld 4812, Australia d Melbourne School of Land and Environment, University of Melbourne, Vic. 3010, Australia b c

A R T I C L E

I N F O

Article history: Received 1 October 2014 Received in revised form 6 March 2015 Accepted 9 March 2015 Available online 30 March 2015 Keywords: Beef cattle Carbon credits C4 grass Grazing Ranch Steers

A B S T R A C T

Leucaena leucocephala (leucaena) is a perennial legume shrub of subtropical regions that has forage characteristics favourable for livestock production, often delivering ruminant liveweight gains that are superior to most other forage systems. Recent work suggests that leucaena mitigates ruminant enteric methane emissions, implying that the shrub may also reduce greenhouse gas (GHG) emissions at the whole farm level. However, the high crude protein content of leucaena relative to endemic grasses can increase livestock urine nitrogen concentration and may increase soil nitrous oxide emissions, potentially offsetting benefits of enteric methane mitigation. Here we examine the effects of leucaena on emissions, production and profitability at the whole farm level by modelling a property in northern Australia, assuming enterprises that specialise in cattle breeding and finishing. To contrast leucaena with a baseline property with Rhodes grass, we modelled three equivalent leucaena scenarios by matching (1) annual average stocking rate, (2) total liveweight production or (3) net farm emissions with that of the baseline, assuming all animals had access to leucaena. To maintain average annual stocking rate or liveweight production, scenarios 1 and 2 carried 5% and 12% less cattle than the baseline because animals on leucaena grew faster and had greater liveweight. In contrast, the number of animals carried and liveweight production in scenario 3 increased by 15% and 31% relative to the baseline, respectively, due to enteric methane abatement and greater liveweight gains. Grazing of leucaena increased soil nitrous oxide emissions by more than 38% in all scenarios, but this did not substantially offset net emissions abatement because nitrous oxide constituted a far smaller proportion of emissions than did methane (<10% and >90%, respectively). In all scenarios, emissions intensity (net farm emissions per unit liveweight sold) caused by grazing leucaena was reduced by more than 23% relative to baseline emissions intensities. This work shows that whilst income from carbon offsets associated with grazing leucaena is small, leucaena has significant potential to increase both animal production and gross margin, whilst reducing emissions intensity. Provided net farm emissions are maintained or reduced, these results suggest that leucaena is conducive to sustainable intensification of beef production in subtropical grazing systems. © 2015 Elsevier Ltd. All rights reserved.

1. Introduction Leucaena leucocephala (leucaena) is a large perennial legume shrub that typically grows with companion pasture grasses in subtropical and tropical regions (Larsen et al., 1998; Radrizzani et al., 2010). Leucaena is long-lived, drought-tolerant and recognised as high value fodder for ruminant livestock due to its palatability and nutritional characteristics, including high crude protein and nonbloating attributes (Jones and Bunch, 2000; Kennedy and Charmley, 2012; McSweeney et al., 2011; Radrizzani et al., 2010). Together these

* Corresponding author. Tel.: +61 3 6430 4501; fax: +61 3 6430 4959. E-mail address: [email protected] (M.T. Harrison). http://dx.doi.org/10.1016/j.agsy.2015.03.003 0308-521X/© 2015 Elsevier Ltd. All rights reserved.

traits deliver ruminant weight gains that are superior to most other tropical forage systems (Aregheore, 1999), and in combination with tropical grasses leucaena pastures are conducive to productive and sustainable livestock farming systems (Shelton and Dalzell, 2007). In northern Australia, there are more than 120,000 animal equivalents (one AE = 450 kg steer/year) grazing 250,000–300,000 ha of leucaena-grass pastures (Michael Burgis, pers. comm., 2015). At the current rates of adoption, the area planted is expected to exceed 500,000 ha by 2017 (Shelton and Dalzell, 2007). Research indicates there is significant potential for reducing cattle enteric methane emissions as the proportion of leucaena in tropical grass–legume mixtures increases (Kennedy and Charmley, 2012). Suppression of enteric methanogenesis is thought to be the mechanism underlying lower methane emissions and may redirect rumen

M.T. Harrison et al./Agricultural Systems 136 (2015) 138–146

fermentation towards other more useful end products (Ouwerkerk et al., 2008), potentially increasing the proportion of energy available for growth or lactation. Together with enhanced liveweight gains these findings suggest a reduction in greenhouse gas (GHG) emissions per unit beef produced (emissions intensity) and sustainable intensification of subtropical beef farming systems. Most previous work on methane mitigation has concentrated on emissions of individual animals, often using measurements from respiration chambers or from in vitro incubations using donor rumen fluid (Kennedy and Charmley, 2012; Meale et al., 2012). There is less information on the effects of leucaena on the emissions intensity of beef production at the whole farm level, particularly when animals are grazing representative mixtures of leucaena and grass pastures, and when animals are considered in realistic cattle breeding and trading scenarios. Indeed, enterprise-scale analyses can reveal important trade-offs caused by management interventions that are not apparent when studied at the animal level. For example, simulations indicate that improving genetic feed-use efficiency (dry matter consumed per unit liveweight gain) can reduce emissions per unit intake and emissions intensity at the animal level, but the feed conserved can facilitate higher stocking rates such that net farm emissions also increase if livestock numbers are increased (Harrison et al., 2014a, 2014b). Analyses at the whole farm level also account for GHGs other than enteric methane, which may form a significant proportion of farm emissions. This is particularly important for leguminous fodder such as leucaena, which can have crude protein contents greater than 25% (Agbede and Aletor, 2004; Bassala et al., 1991; Liu et al., 2010). Since ruminants excrete 75–95% of nitrogen ingested, a high crude protein diet often leads to increased urine nitrogen concentration and, under conditions conducive to denitrification (de Klein and Eckard, 2008), can result in greater soil nitrous oxide emissions (Eckard et al., 2010). The objective of this study was to determine the effects of leucaena on whole farm production, emissions and emissions intensity of a number of trading scenarios for a cattle farming enterprise typical of north central Australia. We also conducted an economic analysis to contrast the relative profitability of leucaena enterprises to those with only pasture grasses, since interventions to livestock farming systems that reduce whole farm emissions intensity may not necessarily reduce economic risk (Ho et al., 2014), and since greater productivity may not translate into greater profitability (Alcock et al., 2014). 2. Materials and methods 2.1. Overview This study used temporal measurements of liveweight (LW) and methane emissions from steers grazing pastures dominated by Rhodes grass (Chloris gayana) or leucaena (L. leucocephala) at two experimental research sites in Queensland, Australia. Pasture nutritional characteristics and livestock diet composition from these experiments were measured using wet chemistry and faecal near infra-red spectroscopy (F.NIRS), respectively. A modelling approach was used to scale measured animal data to the whole farm level in two steps. The first step used measured data from individual animals to generate herd characteristics using an enterprise model, which was then used to parameterise a greenhouse gas (GHG) emissions model. Together these data were used to estimate the net farm emissions of a representative beef enterprise in central Queensland assuming pastures consisting of either Rhodes grass or leucaena hedgerows with interspersed Rhodes grass, consistent with field experiments. The second step examined three scenarios to determine the effects of leucaena on animal productivity, emissions, farm gross margin and the emissions intensity of beef produced and sold.

139

2.2. Measurement of pasture nutritional characteristics, animal diet composition, liveweight and methane emissions: field studies Measured liveweight and methane emissions as well as pasture nutritive characteristics were adopted from field experiments conducted at Belmont Research Station (23.22°S, 150.38°E) near Rockhampton in Queensland, Australia. A concurrent experiment at Brian Pastures (25.64°S, 151.72°E) near Gayndah was used to provide additional data on nutritive characteristics. Field experimentation began in summer 2012/13. At each site, and to coincide with methane measurements, grass pastures were assessed for total available biomass by cutting up to 12 × 0.25 m2 quadrats and species composition determined. Collected grass material was dried at 65 °C to constant weight then ground, bulked and sub-sampled for proximate analysis of dry matter (DM), organic matter (OM), crude protein (CP), and acid- and neutral-detergent fibre content (ADF, NDF). Pluck samples of leucaena leaf and stem (diameter <5 mm) were randomly sampled across each plantation at both sites and dried under similar conditions prior to analysis. Leucaena DM was further sub-sampled and placed onto dry ice, stored at −80 °C and freeze dried for measurement of condensed tannins. Faecal samples from each animal were collected to coincide with methane measurement periods throughout the project at both sites to estimate animal diet composition. Digestibility measures, the proportion of leucaena in the diet and the nutritive quality of the forage consumed were estimated using F.NIRS analysis. Leucaena and grass forage had similar DM and OM content throughout 2013, but the leucaena forage had greater nitrogen and lower fibre concentration compared with the pasture grasses during the dry season (Table 1). Differences in nutritional characteristics were mitigated when diet composition was considered, because diets in the leucaena treatment included both leucaena and grass forage. Over the three sampled periods, the apparent crude protein and in vivo dry matter digestibility of animals grazing leucaena was 2.9–4.3% and 2.6–3.6% greater than that measured for the grasspastured animals, respectively (Table 1). Values in Table 1 were used to parameterise pasture nutritive characteristics and diet composition coefficients in the models for simulating whole farm production and emissions (described below). On Belmont Research Station Bos indicus cross steers (n = 60, initial mean ± SEM LW of 325 ± 7 kg) were allocated to one of two management groups (n = 30); leucaena or grass fed. Animals were transferred to either paddocks of approximately 7 ha containing double rows of L. leucocephala (cv. Cunningham) at 4 m spacing, or paddocks of approximately 6.8 ha dominated by Rhodes grass (C. gayana). At approximately 14 day intervals animals were rotated into adjacent leucaena or Rhodes grass paddocks, respectively, to ensure ad libitum grazing intakes. Grazing of leucaena and Rhodes grass pastures by each group of steers and LW measurements commenced in November 2012 when animals reached 400 days old. Mean daily liveweight gains slowed over time, with the leucaena and grass-fed steers averaging 0.95 and 0.75 kg/head/day respectively for the first 100 days, and 0.59 and 0.46 kg/head/day over 500 days, respectively (Fig. 1). On Brian Pastures B. indicus composite steers (n = 60, initial mean ± SEM LW 237 ± 3.4 kg) were allocated to one of two management groups; leucaena or grass pastures. Animals were transferred to either paddocks containing double rows of leucaena (cv. Cunningham), 4 ha at 3 m spacing, or approximately 15 ha Native Blue grass (Dicanthium spp.) pastures. On 28 day intervals, animals were rotated into adjacent leucaena or grass paddocks, respectively, to ensure ad libitum grazing intakes. Differences in stocking rates across sites reflected regional variation associated with climate and soil characteristics. Open path lasers (GasFinder Boreal Laser Inc., Canada) were used to determine methane emissions at the herd scale for cattle grazing leucaena and grass-pastured groups. This approach measured the

140

M.T. Harrison et al./Agricultural Systems 136 (2015) 138–146

Table 1 Dry matter (DM), organic matter (OM), nitrogen (N), neutral detergent fibre (NDF) and acid detergent fibre (ADF) of available forage diets on Belmont and Brian Pastures Research Stations during each methane measurement campaign (see Fig. 1). Belmont

Brian pastures

Date:

March/April 2013

June/July 2013

Age (days):

465–479

542–591

Forage analyses DM (g/kg) OM (g/kg DM) N (g/kg DM) NDF (g/kg DM) ADF (g/kg DM) Diet composition Diet CP (%) In vivo DMD (%)

March/April 2014 818–846

Pasture

Leucaena

Pasture

Leucaena

Pasture

Leucaena

931 895 11.8 692 389

910 927 60.3 225 189

939 912 8.0 728 458

930 919 38.6 377 247

942 877 3.0 766 588

923 909 36.3 222 160

7.9 56.9

12.0 59.6

8.6 59.4

12.9 63.0

6.1 55.7

9.0 58.3

methane flux relative to the source (cattle) and background emissions for 5 h daily for up to 21 days for each group of animals. Methane measurements on Belmont Research Station were conducted between 23 March and 6 April 2013, between 8 June and 27 July 2013, and between 11 March and 8 April 2014 (Fig. 1). Animal emissions were calculated using backward Lagrangian Stochastic

modelling (Flesch et al., 1995) using WindTrax software (Thunder Beach Scientific, Canada). Emissions were computed on the paddock scale (total g CH4/day) for each group of animals and expressed on an average animal basis (Table 2). Measurements indicated that emissions from cattle with access to leucaena were 21% ± 4% less than those from cattle grazing Rhodes grass only, depending on time of year. The relative difference between methane emissions at each measurement time from cattle groups on either Rhodes grass or leucaena were used to inform the modelling analyses (see below).

2.3. Whole farm modelling of livestock production, gross margin, greenhouse gas emissions and emissions intensities 2.3.1. Overview To scale the animal-level measurements to the whole-farm level, a farm enterprise model was used to generate herd structures and an emissions model was used to estimate GHG, assuming pastures were dominated by Rhodes grass or leucaena, with Rhodes grass representing the baseline. Measurements of liveweight gains, enteric methane emissions, pasture nutritional characteristics and animal diet composition described above were used to parameterise the models. Prices received and selling costs in the modelling scenario analyses were sourced from abattoir slaughter data and recent saleyard reports.

Fig. 1. Mean liveweight of steers on Belmont Research Station grazing leucaena or Rhodes grass pastures (n = 30). Animals entered trial paddocks on 17 January 2013 at 400 days old. Vertical bars represent standard error of the mean and horizontal bars indicate the timing of methane laser measurement periods conducted during 23 March to 6 April 2013, 8 June to 27 July 2013 and 11 March to 8 April 2014.

2.3.2. Pasture and soil characteristics Recent surveys have shown that the majority of leucaena pastures are less than 10 years old and are located in the 600–800 mm/ year rainfall zones of central Queensland (Radrizzani et al., 2010). Around 84% of leucaena pastures in these areas are grown on brigalow clay soils (Vertosols), with mixed leucaena-grass pastures consisting of Buffel grass (Cenchrus ciliaris) and Rhodes grasses, as well as Green and Gatton panic (Panicum maximum spp.; Radrizzani et al., 2010). More than 90% of respondents in these surveys indicated that pastures were rain-fed and did not receive

Table 2 Mean (±SEM) liveweight, average daily gains (ADG) and methane emissions for steers grazing predominantly leucaena or Rhodes grass pastures at Belmont Research Station.

Days of age:

Liveweight (kg/head) ADG (kg/day) CH4 emissions (g/head/day) CH4 emissions (g/LW/day)

March/April 2013

June/July 2013

March/April 2014

465–479

542–591

818–846

Rhodes

Leucaena

Rhodes

Leucaena

Rhodes

Leucaena

384 ± 8 1.1 ± 0.04 210 ± 9.6 0.55

409 ± 8 1.5 ± 0.04 173 ± 15.4 0.42

423 ± 9 0.4 ± 0.02 219 ± 29.8 0.52

460 ± 8 0.7 ± 0.02 143 ± 48.9 0.31

546 ± 10 0.4 ± 0.01 213 ± 34.2 0.39

606 ± 10 0.6 ± 0.01 181 ± 27.9 0.29

M.T. Harrison et al./Agricultural Systems 136 (2015) 138–146

fertiliser (Radrizzani et al., 2010). Following these data the modelling assumed that pastures did not receive irrigation or fertilisation. For consistency with empirical measurements of production, GHG emissions and nutritive data, it was assumed that mixed leucaenagrass pastures consisted of leucaena hedgerows with inter-rows dominated by Rhodes grass, with the proportion of leucaena to grass used in the modelling similar to that measured in the field. Extrapolation of the results shown herein to sites with other grass mixtures would likely depend on the biomass and nutritive differences between Rhodes grass and the alternative grass species, because such characteristics affect animal liveweight gain and whole farm productivity, amongst other factors.

2.3.3. Age- and sex class terminology, herd- and enterprise characteristics Terminologies of different cattle age and sex classes differ widely, particularly on the international scale. In this study we followed terminology used for cattle in northern Australia (Holmes, 2012) and adopted the following definitions throughout: Calves: Heifers: Cows: Breeders: Steers: Bullocks: Bulls:

Animals less than 1 year old Females aged 1 to less than 3 years old Females aged 3 years or greater Females mated in the reproductive herd (either heifers or cows) Castrated males aged 1 to less than 3 years old Castrated males aged 3 years or greater Uncastrated males aged 1 year or greater

Animal numbers and herd structures were generated using Breedcowplus V6.0 (Holmes, 2012; also see https://www.daff .qld.gov.au/business-trade/business-and-trade-services/breedcowand-dynama-software/software-overview). This program is a steadystate model that optimises herd structures annually over a 10 year period according to alternative management and initial numbers of heifers, weaning and culling percentages, mortality rates, and trading of animals in each age classes. Breedcowplus requires inputs for enterprise running- and cattle trading costs, as well as prices received for cattle sold at market, allowing users to contrast the gross margins resulting from alternative trading scenarios, management rules and herd structures. The modelling performed here assumed an enterprise with a self-replacing herd that specialised in beef breeding, growing and finishing, following common practice in central Queensland (AusVet, 2006). Most graziers in the region use leucaena to fatten and finish steers and bullocks, although backgrounding and weaning of cattle on leucaena paddocks is also popular, and many pastures are managed under relatively high grazing pressure (Radrizzani et al., 2010). We adopted the average herd size for the region of 400 adult equivalents (AE) after Dalzell (2005) and tested scenarios in which all animals aged 16 months or greater had access to leucaena (see below). The property size was set at 267 ha to provide a baseline stocking pressure of 1.5 AE/ha after Radrizzani et al. (2010). It was assumed that the proportion of leucaena relative to grass in the animal diet was equal to that measured and that animals did not have access to nutritional supplements, following survey data of graziers in the region (Dalzell, 2005; Eady et al., 2011). Liveweight gains of growing animals pastured on Rhodes grass or leucaena were set to those measured, and liveweight of mature animals over time was assumed constant irrespective of pasture type. To avoid confounding results due to differences in growth rates between sexes and between forage types, it was assumed that heifers weighed 20% less than steers of the same age, following Breedcowplus guidelines (Holmes, 2012). Wherever possible the modelling used measured data, so the requirements of Breedcowplus and the Beef Greenhouse Accounting Framework (BGAF; see below) for liveweights within specific animal

141

Table 3 Liveweight (kg) and age of animals (months of age) in the Rhodes grass or leucaena enterprises. Values were linearly interpolated from measurements taken from cattle grazing each pasture type and were used to calculate annual average stocking rates, profitability, greenhouse gas emissions and emissions intensity in the scenario analyses using Breedcowplus and the Beef Greenhouse Accounting Framework (see text for details). Animal class

Rhodes grass

Leucaena

Heifers 12 Heifers 16 Heifers 24 Cows ≥36 Steers 12 Steers 16 Steers 24 Steers 29 Steers 33 Bulls all ages

209 259 360 520 261 324 450 537 561 800

209 259a 397 520 261 324a 496 600 665 800

a Animals began grazing leucaena at 16 months of age. Prior to this age all animals grazed Rhodes grass.

age bands were estimated by linear interpolation of measured values (Table 3). For consistency with our measured field data, modelled animals were not allowed to graze leucaena until animals reached 16 months of age; prior to this age all animals grazed Rhodes grass (Table 3). Females were first joined (mated) at 12 months of age and every year thereafter until all cows were culled at 11 years of age. The number of mated and unmated females in each age class progressively declined with greater age due to mortality, sales and culling. No differences in animal fertility due to forage type were assumed, with calf weaning rates of 74% for all females mated (or 87% per female mated and retained) for all enterprises. Across all age classes, the majority of unmated females that were not required to maintain the overall property stocking rate were sold at 30 months of age (~50% and 20% of unmated and mated females, respectively). Thereafter, 5% of unmated females and 20% of mated females in each age class were sold annually; all females were sold on the local meat markets. All steers were assumed to be sold as finished animals on the Japanese Oxen (‘Jap Ox’) meat trading market aged 33 months. Bulls were culled and replacements purchased at 6 years of age. All animal mortality rates were set at 2% per annum except for bull mortality rates which were set to 6%. 2.3.4. Costs, prices and gross margin analyses Gross margin analyses were conducted using Breedcowplus assuming an interest payment of 10% of the capital value of the herd. Regional average animal husbandry costs included those for vaccination, replacement breeders and veterinary services (Cullen et al., 2013), assuming best management practice for central Queensland (Table 4). Sale prices received and selling costs shown in Table 5 were sourced from both abattoir slaughter data from steers used in experiments (Teys Australia Ltd.) and from saleyard data

Table 4 Annual husbandry costs per animal retained on farm or sold ($AUD/head) associated with each animal class (years of age). Unit costs for vaccination, veterinary services, replacement breeders etc. were sourced from Cullen et al. (2013). Animal class

Retained

Sold

Calves <1 Heifers 1–2 Heifers 2 to <3 and cows ≥3 Steers 1–2 Steers 2 to <3 Bulls all ages

32.26 27.55 24.90 23.00 24.00 22.31

21.26 14.25 12.00 11.50 12.00 4.50

142

M.T. Harrison et al./Agricultural Systems 136 (2015) 138–146

Table 5 Prices received per unit liveweight, sales commission and other selling costs for each livestock category. Financial metrics are expressed in $AUD and were sourced from sale prices received from abattoir slaughter data and from Meat and Livestock Australia (http://www.mla.com.au/Prices-and-markets). Animal class (years of age)

Price ($/kg LWa)

Commission (% of value)

Other selling costs ($/head)

Freight ($/head)

Female calves <1 Heifers 1–2 Heifers 2 to <3 Cows ≥3–6 Cows 7–10 Male calves <1 Steers 1–2 Steers 2 to <3 Culled bulls all ages

1.61 1.60 1.60 1.35 1.30 1.99 1.83 1.80 1.25

5.0 2.5 0.0 0.0 0.0 2.5 0.0 0.0 0.0

12.00 10.00 5.00 5.00 5.00 12.00 10.00 5.00 5.00

8.57 12.00 22.50 24.55 24.55 8.57 12.86 27.00 27.00

a Liveweight at sale differed according to enterprise and scenario (see Table 3 and section 2.3.5).

published by Meat and Livestock Australia (http://www.mla.com. au/Prices-and-markets). Prices per unit liveweight of leucaena and Rhodes grass animals were assigned the same values following abattoir data, which did not reveal significant differences in carcase attributes between treatments. Carbon offset income from Australian Carbon Credit Units (ACCU) under a scheme such as the Carbon Farming Initiative (CFI) or Emissions Reduction Fund (ERF) was included where appropriate, with emissions reductions of leucaena enterprises calculated relative to the emissions from the herd grazing the Rhodes grass pastures (deemed as the baseline). Each tonne of CO2–equivalent (CO2-e) emissions was valued at either $10 or $23 to represent current and historical prices and to provide an indication of gross margin sensitivity to carbon price. This analysis compared the gross margin and emissions of steady-state enterprises (costs and emissions associated with planting and establishment of leucaena were not included). Costs associated with CFI participation and compliance were valued at $2000 per year following Davison and Keogh (2011), assuming leucaena farmers would complete an annual statement of activities performed, compliance with methodologies for methane mitigation, as well as any changes made during the project period. 2.3.5. Enterprise scenario analyses The productivity, gross margin and GHG emissions of a baseline farm with Rhodes grass pastures was compared with that of an equivalent farm with leucaena under three scenarios to encompass the variation in beef trading across central Queensland. Scenarios 1, 2 and 3 benchmarked leucaena enterprises against the baseline with equal annual average stocking rate (400 AE), total liveweight production (74.3 t LW/year) or net farm emissions (734 t CO2-e) respectively. Steers in scenarios 1–3 were sold in their third year early in the dry season (June) aged 33 months, typical of cattle sold on the Jap Ox market in northern Australia. For consistency with field data, it was assumed that animals in the leucaena enterprises were not given access to the shrubs until 16 months of age. After this age it was assumed that all growing and mature animals in scenarios 1–3 had (1) the enteric methane mitigation, (2) the diet composition and intake and (3) the liveweight gains (until maturity when constant liveweight was assumed) that was measured in the field experiments. Each scenario assumed sufficient forage availability to support changes in stocking rate such that dry matter intake was unrestricted and liveweight gains were not affected by grazing pressure or feed limitation. This assumption was based on measured data and was necessary to prevent confounding results due to changes in stocking rate with those due to changes in the relative proportion of animals in each age class across scenarios.

2.3.6. Whole farm greenhouse gas emissions and emissions intensities Whole farm greenhouse gas (GHG) emissions in tonnes of CO2 equivalents were estimated using the Beef Greenhouse Accounting Framework (B-GAF; see http://www.greenhouse.unimelb. edu.au/Tools.htm and Browne et al., 2011) which uses Australian National Greenhouse Inventory methods prescribed by the DCCEE (2012). Greenhouse gas emissions included those from livestock (enteric and manure), urine, dung and indirect emissions from ammonia volatilisation as well as organic nitrogen leaching and runoff. Whole-farm emissions intensities were computed as the dividend of annual net farm GHG emissions and annual meat sales from all stock (t CO2-e/t LW sold). To determine the variation in the carbon footprint across livestock classes, emissions per animal or per unit liveweight were estimated for animals sold or retained annually on farm (carried) using only livestock emissions from each animal class. In contrast, emissions intensity was determined using whole farm emissions and total animal production (LW sold). 2.3.7. Sensitivity analysis The sensitivity of liveweight production, emissions intensity and gross margin per adult equivalent to ±10% variation in key inputs (liveweight gain, relative methane mitigation, crude protein and dry matter digestibility of forage, prices received and enterprise costs) was examined by determining the percentage change in each output relative to its original value. 3. Results 3.1. Scenario 1: leucaena enterprises with equivalent stocking rates to the baseline Allowing livestock to graze leucaena increased animal growth rates and liveweight such that the number of cattle carried on farm was reduced in order to maintain stocking rate (Table 6). Even though this reduced total annual animal sales by 5%, total liveweight production and profitability increased by 8% and 17%, respectively. When carbon was valued at $23 t/CO2-e potential carbon offset income from farmer participation in a scheme such as the CFI contributed ~$2900 to an additional ~$13,000 gross margin received from a leucaena enterprise run at the same stocking rate as the Rhodes grass enterprise (scenario 1, Table 6). The higher crude protein content of leucaena increased nitrous oxide production by 49% relative to Rhodes grass in scenario 1, but this did not substantially offset a 17% reduction in net emissions because nitrous oxide constituted a much smaller proportion of whole farm emissions than did methane (~5% and ~95%, respectively; Table 6 and Fig. 2). In both cases, nitrous oxide from dung and urine was around double that from indirect sources including nitrogen lost as ammonia volatilisation (Fig. 2). 3.2. Scenario 2: leucaena enterprises with equivalent liveweight production to the baseline When total annual liveweight production of the leucaena enterprise was matched with that of the baseline, stocking rate and total livestock sales decreased by 8% and 12%, respectively (scenario 2, Table 6). This resulted in a carbon offset (CFI) income of ~$4000 at $23 t/CO-e, though gross margin was around $6000 less than leucaena enterprises run at equivalent stocking rates (cf. scenarios 1 and 2 in Table 6). Scenario 2 had the lowest net farm emissions of all leucaena scenarios examined, with reductions in total and methane emissions of more than 24%, although nitrous

M.T. Harrison et al./Agricultural Systems 136 (2015) 138–146

143

Table 6 Annual production, gross margin and greenhouse gas emissions of three alternative leucaena enterprises in scenarios with equal numbers of adult equivalents (AE), equal total liveweight production (LW) or equal net farm emissions to that of the baseline Rhodes grass enterprise. Carbon offset income from greenhouse gas emissions abatement under a scheme such as the Carbon Farming Initiative (CFI) was computed relative to the baseline assuming carbon prices of $23 or $10/t CO2-e. Values in parentheses indicate percentage change relative to the baseline. Baseline Production Total adult equivalents (AE) Total cattle carried (head) Total breeders mated (head) Total cows and heifers sold (head) Total steers and bullocks sold (head) Total animals sold (head) Total liveweight production (t LW) Economicsa CFI income @ $23/t CO2-e ($) CFI income @ $10/t CO2-e ($) Gross margin with CFI @ $23/t ($) Gross margin per adult equivalent ($/AE) Greenhouse gas emissions Total methane emissions (t CO2-e/farm) Total nitrous oxide emissions (t CO2-e/farm) Net farm GHGb emissions (t CO2-e/farm) Emissions per animal sold (t CO2-e/head) Emissions per animal carried (t CO2-e/head) Emissions per LW sold (t CO2-e/t LW sold) Emissions per LW carried (t CO2-e/t LW) Emissions intensity (t CO2-e/t LW sold) a b

400 406 205 71 73 145 75.1 0 0 79,647 199 699 38 738 2.1 1.6 4.1 4.7 9.8

Scenario 1: leucaena equal AE 400 (0) 385 (−5) 194 (−5) 67 (−5) 69 (−5) 138 (−5) 81.4 (8) 2944 1280 93,021 (17) 233 (17) 553 (−21) 57 (49) 610 (−17) 1.9 (−9) 1.4 (−13) 3.3 (−21) 3.9 (−18) 7.5 (−24)

Scenario 2: leucaena equal LW 369 (−8) 355 (−12) 179 (−12) 62 (−12) 64 (−12) 127 (−12) 75.2 (0) 4018 1747 86,959 (9) 236 (18) 511 (−27) 52 (38) 563 (−24) 1.9 (−9) 1.4 (−13) 3.3 (−21) 3.9 (−18) 7.5 (−24)

Scenario 3: leucaena equal emissions 483 (21) 465 (15) 235 (15) 81 (15) 84 (15) 166 (14) 98.2 (31) 0 0 109,166 (37) 226 (14) 670 (−4) 68 (81) 738 (0) 1.9 (−9) 1.4 (−12) 3.3 (−21) 3.9 (−18) 7.5 (−23)

$AUD. Greenhouse gas.

oxide emissions were still 38% greater than those of the Rhodes grass enterprise (Table 6, Fig. 2). 3.3. Scenario 3: leucaena enterprises with equivalent net farm emissions to the baseline To match net farm emissions of the leucaena enterprise with that of the baseline (scenario 3), stocking rate increased by 21%, resulting in a 14% increase in animal sales. Liveweight sales and gross margin increased even further (31% and 37% greater than the baseline, respectively), notwithstanding the fact that there was no carbon offset income (Table 6). In scenario 3 methane emissions were around 4% lower than baseline methane production and nitrous oxide emissions were around 81% greater, indicating the relative change in the emissions profile caused by introducing leucaena when net farm emissions are maintained (Table 6, Fig. 2). 3.4. Average annual production, emissions and emissions intensity of each animal class

Fig. 2. Net farm greenhouse gas emissions and liveweight sales of a baseline beef enterprise with Rhodes grass pastures and three equivalent enterprises with leucaena pastures. Left-hand bars within each pair represent methane and nitrous oxide emissions. Right-hand bars within each scenario show total liveweight sales and the relative contribution of each age and sex. Labels on the horizontal axis represent emissions from the baseline enterprise and leucaena enterprises with equal adult equivalents (scenario 1), liveweight turnoff (scenario 2) or net farm emissions (scenario 3) to the baseline. Note that the greenhouse gas emissions are dominated by enteric methane wherein the left axis does not begin at the origin.

Steers aged 2–3 years and cows constituted the majority of emissions and liveweight carried on farm, with each accounting for 24–26% of liveweight carried and 22–24% of emissions from animals carried on farm (Fig. 3). There was a tendency for leucaena enterprises to shift total liveweight and emissions from animals carried away from cows and towards steers, but these effects were relatively small. The metric used to compute emissions intensity strongly influenced the relative differences between animal classes. For instance, calves had the greatest emissions per unit liveweight carried but the lowest emissions per animal, whereas bulls had the lowest emissions per unit liveweight carried and the highest emissions per animal (Fig. 4). Irrespective of the metric used for computation, emissions intensities of all animal classes except calves were lower for enterprises that allowed livestock access to leucaena (there were no differences between emissions intensities of calves because young animals were not allowed access to leucaena until 16 months of age).

144

M.T. Harrison et al./Agricultural Systems 136 (2015) 138–146

Fig. 3. Net greenhouse gas emissions from animals carried and liveweight carried of the baseline beef enterprise with Rhodes grass pastures and three equivalent enterprises with leucaena pastures. Left- and right-hand bars within each pair represent emissions and liveweight of each animal class carried on farm. Horizontal axis labels are defined in Fig. 2.

3.5. Sensitivity analysis The sensitivity of total liveweight production, emissions intensity and gross margin per adult equivalent was examined via ±10% variation in each of the main inputs (Table 7). A preliminary analysis revealed that the sensitivity of most outputs of the baseline and leucaena scenarios were similar, so results shown here relate only to the baseline and leucaena scenario 1. Total liveweight produced and emissions intensity were relatively insensitive to perturbation of measured variables (<3% variation). Gross margin per adult equivalent was more sensitive to reductions rather than increases in liveweight gain, suggesting diminishing returns per adult equivalent as liveweight gain increased. Gross margin per adult equivalent was most sensitive to prices received and was more than twice as sensitive to prices as to costs, because prices were attributed per unit liveweight whereas costs were mostly attributed per unit animal. 4. Discussion 4.1. Leucaena improved whole farm production and profitability

Fig. 4. Greenhouse gas emissions per unit liveweight (a) and per head (b) of animals carried annually in the baseline beef enterprise with Rhodes grass pastures and three equivalent enterprises with leucaena pastures. Horizontal axis labels are defined in Fig. 2.

The primary aim of this work was to examine the effects of leucaena on whole farm production, GHG emissions and emissions intensity, and a secondary aim was to examine how leucaena affected gross margin. We have shown that leucaena improved total liveweight sales by 8% and 31% in scenarios 1 and 3 respectively, and that leucaena reduced net farm emissions by 17% and 24% in scenarios 1 and 2 respectively. Together these factors accounted for a 23% reduction in whole farm emissions intensity, indicating that leucaena has clear benefits in terms of reducing farm carbon footprint. In all cases, enterprises with leucaena were more profitable than those with grass-only pastures, although relative profitability strongly depended on the scenario examined. The greatest profit was obtained in scenarios that matched net farm emissions with those of the baseline by increasing stocking rates; higher profit was elicited by improved diet quality and enhanced liveweight gains of

M.T. Harrison et al./Agricultural Systems 136 (2015) 138–146

Table 7 Sensitivity of total liveweight production, emissions intensity and gross margin per adult equivalent (AE) to ± 10% variation in key input variables for the baseline and for leucaena scenario 1 (see Table 6 for descriptions). Values in parenthesis represent percentage change relative to actual values.

Total liveweight production (t LW sold) Actual value −10% LW gain +10% LW gain Emissions intensity (t CO2e/t LW sold) Actual value −10% LW gain +10% LW gain −10% methane mitigation +10% methane mitigation −10% crude protein +10% crude protein Gross margin per adult equivalent ($/AE) Actual value −10% LW gain +10% LW gain −10% methane mitigation +10% methane mitigation −10% crude protein +10% crude protein −10% dry matter digestibility +10% dry matter digestibility −10% prices received +10% prices received −10% costs +10% costs

Baseline

Leucaena equal AE

75.1 74.0 (−1.5) 76.8 (2.2)

81.4 80.3 (−1.4) 82.5 (1.3)

9.8 9.9 (1.0) 9.7 (−1.6) 9.8 (0.0) 9.8 (0.0) 9.8 (−0.5) 9.9 (0.5)

7.5 7.6 (1.5) 7.3 (−2.2) 7.3 (−2.3) 7.7 (2.3) 7.3 (−2.3) 7.6 (0.9)

199 189 (−5.3) 202 (1.5) 199 (0) 199 (0) 199 (0) 199 (0) 199 (0)

233 222 (−4.3) 242 (3.9) 233 (0.3) 232 (−0.3) 233 (0.3) 232 (−0.1) 233 (0)

199 (0)

233 (0)

174 (−12.7) 225 (12.8) 210 (5.5) 188 (−5.5)

204 (−12.1) 261 (12.2) 244 (4.8) 221 (−4.8)

Source: Cullen et al. (2013).

animals grazing leucaena. Scenario 3 had 31% greater annual liveweight production than that of the baseline and 37% higher gross margin (Table 6). Scenario 3 was devised on the premise of a future legislated cap on net farm emissions, with the increase in stocking rates justified by previous work suggesting that leucaena pastures can be grazed at significantly higher stocking rates than grass pastures (Shelton and Dalzell, 2007). Since gross margins were computed using fixed costs and prices and since prices were shown to be a relatively sensitive input (Table 7), future economic studies should account for additional factors that were beyond the scope of the current study, such as costs associated with planting leucaena and the time required to recover initial capital outlays. Given that our modelling used experimental data measured over two years, a future iteration on this work might be to use a dynamic model to simulate leucaena production over several years. We expect that the use of a dynamic model would reveal important insights into seasonal production due to variability in the weather, particularly since leucaena persists through drought, but is intolerant of heavy frosts (Shelton and Brewbaker, 1998). A dynamic model could also provide insight into how herd numbers and pasture nutritive characteristics vary seasonally in response to climate, changes in biomass associated with growth or soil carbon emissions caused by cultivation at planting. 4.2. Leucaena significantly reduced whole farm emissions, primarily due to mitigation of enteric methane A concern with respect to whole farm emissions associated with leucaena might be increased nitrous oxide emissions due to the relatively high crude protein content and increased nitrogen cycling through animals. We have shown that grazing of leucaena pastures

145

was conducive to significant increases in nitrous oxide production, ranging from 38% to 81% in the scenarios examined (14–30 t CO 2 -e). Higher urine nitrogen concentration led to increased ammonia volatilisation which further increased nitrous oxide emissions, albeit to a lesser extent than that emitted from animal excrement directly (Fig. 2). Changes in nitrous oxide production had minimal effects on emissions at the whole farm level because enteric methane emissions were an order of magnitude greater (Table 6). Under scenarios of equal adult equivalents or liveweight production, leucaena reduced whole farm emissions by more than 17% due to its anti-methanogenic properties. In circumstances of drought where livestock intake of grasses is reduced due to limited dry matter availability and causes luxury intake of leucaena, higher dietary crude protein content would have little impact on overall farm emissions because the additional nitrous oxide emissions would be small relative to the mitigation of enteric methane (Tables 6 and 7). These observations concur with past studies documenting emissions of other livestock systems (Alcock et al., 2014; Harrison et al., 2014a, 2014b) and highlight the importance of targeting enteric methane in emissions mitigation strategies, since enteric methane dominates the farm emissions profile (Fig. 2). 4.3. Leucaena improved whole farm emissions intensity Concurrent benefits to productivity and to methane emissions mitigation attributed to leucaena reduced emissions intensities by 23–24%, underscoring the advantages associated with the legume to farmers, the beef industry and the environment. In comparing emissions intensity across industries these results substantiate the importance of contrasting the metric in the same dimensions, because reductions in emissions per unit product sold differ from those per animal sold (Table 6) or per animal retained on farm (Fig. 4). Indeed, when contrast per unit liveweight carried calves had the highest emissions intensities, but when contrast per head calves had the lowest emissions intensity, and bulls were the converse (Fig. 4). Such differences concord with results of Charmley et al. (2008) for another emissions intensity metric, which showed that male calves (compared with bulls) had the lowest (highest) methane emissions per unit liveweight gain. 4.4. Carbon offset income was small relative to income from increased animal production Enterprises with leucaena that had equivalent stocking rates or total liveweight production (scenarios 1 and 2) to comparable grass enterprises had lower net farm emissions, resulting in carbon offset income (e.g. Carbon Farming Initiative or Emissions Reduction Fund) of ~$2900 and ~$4000, respectively (when Australian Carbon Credit Units were valued at $23/t CO2-e; Table 6). Together with increased productivity these scenarios increased profitability by 9–18%. Such financial gains were less than half the additional gross margin realised by maintaining baseline levels of farm emissions and increasing liveweight production on leucaena pastures (scenario 3). These results imply that graziers using leucaena in beef enterprises would receive higher financial returns by using leucaena to increase production and foregoing potential GHG emissions mitigation income. Collectively these findings are not surprising given that – at current prices – an average tonne of beef is worth 69– 159 times more than a tonne of CO2–e ($1592/t LWT cf. $23/t CO2-e or $10/t CO2-e; Table 5). These differences highlight the need for substantial increases in current carbon pricing mechanisms associated with GHG abatement if such schemes are likely to become financially attractive to farmers, and align with other analysis of carbon mitigation income resulting from interventions to livestock farming systems (Harrison et al., 2014a, 2014b; Ho et al., 2014).

146

M.T. Harrison et al./Agricultural Systems 136 (2015) 138–146

5. Conclusions Irrespective of whether enterprises conducting leucaena grazing had the same number of adult equivalents, total liveweight production or net farm emissions compared with enterprises with subtropical pasture grasses, leucaena’s superior nutritive characteristics and anti-methanogenic properties reduced emissions intensity by at least 23%. The relatively high crude protein content of leucaena increased emissions of nitrous oxide, but since enteric methane emissions were around an order of magnitude greater, effects of increased nitrous oxide on net farm emissions were small. Provided that leucaena is grown sufficiently densely to support greater stocking rates, this study has shown that financial returns resulting from greater animal production and maintenance of net emissions under leucaena (scenario 3) are much greater than returns from mitigating greenhouse gas emissions and maintaining current stocking rates or total liveweight production (scenarios 1 and 2) under a government scheme crediting farmers for carbon emissions mitigation.

Acknowledgements This project was supported by The University of Melbourne, The Tasmanian Institute of Agriculture and CSIRO through funding from the Australian Government Department of Agriculture, Dairy Australia, Meat and Livestock Australia and Australian Wool Innovation. We thank Michael Burgis and The Leucaena Network for valuable information on current practices conducted by graziers using leucaena in northern Australia.

References Agbede, J.O., Aletor, V.A., 2004. Chemical characterization and protein quality evaluation of leaf protein concentrates from Glyricidia sepium and Leucaena leucocephala. Int. J. Food Sci. and Tech. 39, 253–261. Alcock, D.J., Harrison, M.T., Rawnsley, R.P., Eckard, R.J., 2015. Can animal genetics and flock management be used to reduce greenhouse gas emissions but also maintain productivity of wool-producing enterprises? Agric. Syst. 132, 25–34. doi:10.1016/ j.agsy.2014.06.007. Aregheore, E.M., 1999. Nutritive and anti-nutritive value of some tree legumes used in ruminant livestock nutrition in Pacific Island countries. J. South Pac. Agric. 6, 50–61. AusVet, 2006. A Report to the Australian Department of Agriculture, Fisheries and Forestry by AusVet Animal Health Services: Structure and Dynamics of the Australian Beef Cattle Industry. (accessed 05.05.14). Bassala, P., Felker, P., Swakon, D.H.D., 1991. A comparison of Leucaena leucocephala and Leucaena pulverulenta leaf and stem age classes for nutritional value. Trop. Grassl. 25, 313–316. Browne, N.A., Eckard, R.J., Behrendt, R., Kingwell, R.S., 2011. A comparative analysis of on-farm greenhouse gas emissions from agricultural enterprises in south eastern Australia. Anim. Feed Sci. Technol. 166–67, 641–652. Charmley, E., Stephens, M.L., Kennedy, P.M., 2008. Predicting livestock productivity and methane emissions in northern Australia: development of a bio-economic modelling approach. Aust. J. Exp. Agric. 48, 109–113.

Cullen, B.R., Timms, M., Eckard, R.J., Mitchell, R.A., Whip, P., Phelps, D., 2013. The effect of earlier mating and improving fertility on emissions intensity of beef production in a northern Australian herd. In “Proceedings of the 5th Greenhouse gases and animal agriculture conference”. 23–26 June, Dublin Ireland. Advances in Animal Biosciences 4, 403. de Klein, C.A.M., Eckard, R.J., 2008. Targeted technologies for nitrous oxide abatement from animal agriculture. Aust. J. Exp. Agric. 48, 14–20. Dalzell, S., 2005. An investigation into the extent and causes of leucaena toxicity in Queensland. Final Report to Meat and Livestock Australia, May 2005, 1–42. Davison, S., Keogh, M., 2011. The implications of the Australian Government’s Carbon Farming Initiative for beef producers. Research Report. Australian Farm Institute, Surry Hills, Australia. DCCEE, 2012. National Inventory Report 2010: The Australian Government Submission to the UN Framework Convention on Climate Change April 2012, vol. 1. Department of Climate Change and Energy Efficiency, Canberra, ACT, Australia. Eady, S., Viner, J., MacDonnell, J., 2011. On-farm greenhouse gas emissions and water use: case studies in the Queensland beef industry. Anim. Prod. Sci. 51, 667–681. Eckard, R.J., Grainger, C., de Klein, C.A.M., 2010. Options for the abatement of methane and nitrous oxide from ruminant production: a review. Livest. Sci. 130, 47–56. Flesch, T.K., Wilson, J.D., Yee, E., 1995. Backward-time Lagrangian stochastic dispersion models and their application to estimate gaseous emissions. J. Appl. Meteorol. 34, 1320–1332. Harrison, M.T., Christie, K.M., Rawnsley, R.P., Eckard, R.J., 2014a. Modelling pasture management and livestock genotype interventions to improve whole-farm productivity and reduce greenhouse gas emissions intensities. Anim. Prod. Sci. 54, 2018–2028. http://dx.doi.org/10.1071/AN14421. Harrison, M.T., Jackson, T., Cullen, B.R., Rawnsley, R.P., Ho, C., Cummins, L., et al., 2014b. Increasing ewe genetic fecundity improves whole-farm production and reduces greenhouse gas emissions intensities: 1. Sheep production and emissions intensities. Agric. Syst. 131, 23–33. http://doi:10.1016/j.agsy.2014.07.008. Ho, C., Jackson, T., Harrison, M.T., Eckard, R.J., 2014. Increasing ewe genetic fecundity improves whole-farm production and reduces greenhouse gas emissions intensities: 2. Economic performance. Anim. Prod. Sci. 54, 1248–1253. http:// dx.doi.org/10.1071/AN14309. Holmes, W.E., 2012. Breedcowplus Software Package, Version 6.0. Queensland Department of Primary Industries and Fisheries. Jones, R.M., Bunch, G.A., 2000. A further note on the survival of plants of Leucaena leucocephala in grazed stands. Trop. Agric. 77, 109–110. Kennedy, P.M., Charmley, E., 2012. Methane yields from Brahman cattle fed tropical grasses and legumes. Anim. Prod. Sci. 52, 225–239. Larsen, P.H., Middleton, C.H., Bolam, M.J., Chamberlain, J., 1998. Leucaena in large-scale grazing systems: challenges for development. In: Shelton, H.M., Gutteridge, R.C., Mullen, B.F., Bray, R.A. (Eds.), Leucaena – Adaptation, Quality and Farming Systems. Australian Centre for International Agricultural Research, Canberra, pp. 324–330. Liu, H., Li, J., Duan, Y., Qiu, Q., 2010. Analysis and evaluation on nutrition components of Leucaena leucephala leaves. Guizhou Agric. Sci. 144–145. McSweeney, C.S., Ngu, N.T., Halliday, M.J., Graham, S.R., Giles, H.E., Dalzell, S.A., et al., 2011. Enhanced ruminant production from leucaena – new insights into the role of the ‘leucaena bug’. SAADC 2011 strategies and challenges for sustainable animal agriculture-crop systems, Volume I: invited papers. Proceedings of the 3rd International Conference on sustainable animal agriculture for developing countries, Nakhon Ratchasima, Thailand, 26–29 July, 2011. Meale, S.J., Chaves, A.V., Baah, J., McAllister, T.A., 2012. Methane production of different forages in in vitro ruminal fermentation. Asian-Aust. J. Anim. Sci. 25, 86–91. http://dx.doi.org/10.5713/ajas.2011.11249. Ouwerkerk, D., Turner, A.F., Klieve, A.V., 2008. Diversity of methanogens in ruminants in Queensland. Aust. J. Exp. Agric. 48, 722–725. Radrizzani, A., Dalzell, S.A., Kravchuk, O., Shelton, H.M., 2010. A grazier survey of the long-term productivity of leucaena (Leucaena leucocephala)-grass pastures in Queensland. Anim. Prod. Sci. 50, 105–113. Shelton, H.M., Brewbaker, J.L., 1998. Leucaena leucocephala – the most widely used forage tree legume. In: Gutteridge, R.C., Shelton, H.M. (Eds.), Forage Tree Legumes in Tropical Agriculture. Department of Agriculture, The University of Queensland, Queensland, Australia. (Chapter 2) . Shelton, M., Dalzell, S., 2007. Production, economic and environmental benefits of leucaena pastures. Trop. Grassl. 41, 174–190.